Active Learning: Problem Settings and Recent Developments

By Hideitsu Hino et al

Published on Dec. 16, 2020

Read the original document by opening this link in a new tab.

Table of Contents

Abstract
1 Introduction
2 Problem Setting
2.1 Stream-based active learning
2.2 Pool-based active learning
3 Acquisition Function
3.1 Design of acquisition function

Summary

Active Learning: Problem Settings and Recent Developments by Hideitsu Hino et al discusses the method of obtaining predictive models with high precision at a limited cost through the adaptive selection of samples for labeling. It explains the basic problem settings of active learning and recent research trends, including research on learning acquisition functions, theoretical work on active learning algorithms, and stopping criteria for sequential data acquisition. Application examples for material development and measurement are also introduced. The paper further delves into the concept of stream-based and pool-based active learning, highlighting the importance of the acquisition function in determining which samples require labeling. Various design aspects and methodologies for active learning are discussed, emphasizing the significance of selecting informative and representative samples for efficient learning.